Hyperpolarized MRI of Human Prostate Cancer Reveals Increased Lactate with Tumor Grade Driven by Monocarboxylate Transporter 1

Abstract

Metabolic imaging using hyperpolarized magnetic resonance can increase the sensitivity of MRI, though its ability to inform on relevant changes to biochemistry in humans remains unclear. In this work, we image pyruvate metabolism in patients, assessing the reproducibility of delivery and conversion in the setting of primary prostate cancer. We show that the time to max of pyruvate does not vary significantly within patients undergoing two separate injections or across patients. Furthermore, we show that lactate increases with Gleason grade. RNA sequencing data demonstrate a significant increase in the predominant pyruvate uptake transporter, monocarboxylate transporter 1. Increased protein expression was also observed in regions of high lactate signal, implicating it as the driver of lactate signal in vivo. Targeted DNA sequencing for actionable mutations revealed the highest lactate occurred in patients with PTEN loss. This work identifies a potential link between actionable genomic alterations and metabolic information derived from hyperpolarized pyruvate MRI.

Trial registration: ClinicalTrials.gov NCT02421380.

Keywords: glycolytic flux; hyperpolarized pyruvate; in vivo kinetics; metabolic imaging.

Conflict of interest statement

Declaration of Interests: APC and FR are employed by General Electric Healthcare. JT is employed by Berkshire Magnetics. HH is on the board of IBA. KRK serves on the SAB of NVision Imaging Technologies.

Copyright © 2019 Elsevier Inc. All rights reserved.

Figures

Figure 1.. Hyperpolarization and acquisition

The [1-13C] pyruvate is polarized using dissolution dynamic nuclear polarization: microwave irradiation is used to transfer the high polarization of the radical achieved at high magnetic field and low temperature to the 13C nuclei (red circles). Following dissolution, the sample must pass a Quality Control check before injection, which measures pyruvate and residual radical concentrations, temperature, pH, and polarization. A power injector is used to deliver a bolus of pyruvate with a saline flush. An EPSI sequence is repeated to acquire 2D dynamic spectra with 1.5 cm3 nominal spatial resolution and 4.9 s temporal resolution.

Figure 2.. Data processing

(A) Array of spectra overlaid on a T2-weighted image shows HP signal localized in the prostate. (B) Early and late spectra from one of the voxels in the right peripheral zone show pyruvate delivery and subsequent conversion to lactate. (C) Spectra are automatically fit with Lorentzian curves, which are used to calculate the area of the pyruvate and lactate peaks for each voxel (in space and time). (D) Pyruvate and lactate curves for the whole prostate show delivery of pyruvate and its conversion to lactate. (E) The maximum lactate to total carbon ratio (Lacmax) is shown and was used as the primary metric for analysis. A pathologist marked regions of Primary Gleason grades 3 and 4/5 on histology slides.

Figure 3.. Quantified Metabolic Dynamics and Comparison to Primary Gleason Grade in Prostate Cancer Patients

(A) Pyruvate delivery to the prostate varies across patients (N=16 patient injections) as measured from the time of injection (time-to-max standard deviation = 5.4 s). When corrected for delivery, the delivery curves align, both in shape and time-to-max (standard deviation = 2.4 s). (B) The range of time-to-max decreases for both pyruvate (22.3 s to 6.2 s) and lactate (21.4 s to 14.3 s) when corrected for the input timing (N=16 patient injections). The time-to-max pyruvate shows a bimodal distribution with a separation of 4.9 s, corresponding to the temporal resolution of the acquisition. (C) There is no significant difference in time-to-max pyruvate and lactate between injections in the same patient (p=0.24,0.78, N=5 patients). 4 of the 5 patients were injected and imaged again 1 hr after the first injection, while 1 patient (labeled in blue) was injected and imaged for the second time on a different day. (D) The maximum lactate ratio (Lacmax) maps show similar spatial distributions to the regions marked on histology slides by a pathologist. Furthermore, regions of high-grade tumor (Gleason grade 4/5) show higher ratios than those of low-grade tumor (Gleason grade 3). (E) The Lacmax increases with tumor Gleason grade. The ratios are significantly lower in normal tissue than in tumors (p=0.0001 for Gleason grade 3, p<0.0001 for Gleason grade ≥4). (F) There is no significant difference in Lacmax between test/re-test injections within individual patients (N=5 patients, 2 injections per patient). All plotted data is represented as mean and SD. Significance was tested using a 2-sided Student’s T-test and p-values < 0.05 were considered signficant.

Figure 4.

(A) Schematic for conversion of HP pyruvate to lactate and potential mediators of this reaction in vivo (B) Analysis of RNA sequencing data from the TCGA (Cancer Genome Atlas Research, 2015) showing elevated MCT1 in regions of prostate cancer as compared to regions of normal prostate tissue in the same patient plotted as mean and SD. (C) Representative images from MCT1 immunohistochemistry in matched benign and tumor regions of 7 patients demonstrating an increase in MCT1 in tumor regions. White scale bar represents 500 μm. (D) T2-weighted anatomic MRI with representative regions of interest and corresponding dynamics of HP lactate and pyruvate. The regions of green, red and orange correspond to the ROIs on the anatomic MRI. (E) MCT1 immunohistochemistry for the corresponding regions in (D) showing higher MCT1 staining in the orange and red regions, both of which were Gleason grade 4/5 as confirmed by H&E staining. Scale bar represents 500 μm. (F) Lacmax is shown for each ROI quantified. 5 ROIs had loss of PTEN (red) and 4 demonstrated TMPRSS-ERG fusion (blue) as measured by IMPACT sequencing (G) The average Lacmax is shown for PTEN wildtype (wt, n=11 ROIs) versus PTEN loss (n=5 ROIs, 2 homozygous loss and 3 heterozygous loss) with a P=0.059. (H) The average Lacmax is shown for TMPRSS-ERG fusion (n=4 ROIs) versus wildtype (wt, n=12 ROIs) with a P=0.730. Significance was tested for all data using a 2-sided Student’s T-test and p-values are reported for each comparison.